Releases: Ikomia-dev/IkomiaStudio
Releases · Ikomia-dev/IkomiaStudio
0.12.0
New features:
- Feature to enable/disable workflow task for running
- VTK 3D viewer
- New HUB algorithms:
- infer_ddcolor_colorisation
- infer_owl_v2
- infer_yolo_v9_seg, train_yolo_v9_seg
- infer_depth_anything_v2
- infer_florence_2_caption, infer_florence_2_ocr, infer_florence_2_detection, infer_florence_2_segmentation
- infer_segment_anything_2
- infer_rembg
- infer_flux_1 (schnell)
- infer_finegrain_segmenter
- infer_donut
Improvements:
- Manage EXIF metadata (orientation)
- Content-type header in http request
Bug fixes:
- Color mismatch in semantic segmentation tasks
- Class labels overflow in semantic segmentation IO
- Unhandled exception when reloading erroneous plugin
0.11.1
v0.11.0
New features:
- Embedded interpreter upgrade to Python 3.10
- OpenCV 4.8.1
- CUDA 11.8
- Add feature to edit workflow information
- Add feature to install algorithm requirements manually (#28)
- Add feature to expose task parameters at workflow level
- Add feature to expose task outputs at workflow level
Improvements:
- Keep current parameters values when reloading algorithms in a workflow
- Algorithm documentation window
- Zoom sensitivity in image view
Bug fixes:
v0.10.0
This release marks the transition to the new Ikomia Scale platform. This platform includes a new version of Ikomia HUB with more insights on how to use every algorithms with this API.
New features:
- Switch to new Ikomia HUB from Ikomia Scale platform
- Add get algorithms from public and private HUBs
- Add possibility to push algorithms to private HUB
- Add to publish private algorithms to public HUB
- Add possibility to push workflows to Ikomia Scale
- Token-based authentication for private HUB
Improvements:
- Add polygons output for instance and semantic segmentation algorithms
- Improve speed of auto-completion system (ik module)
- Algorithm template includes a markdown readme
- Display markdown readme in algorithm details
Bug fixes:
- Missing Python binding for CInstanceSegIO::is_data_available()
- Windows: fix download function with correct path separators
- Avoid creating MLFlow experiment in training algorithm constructor
- Paint issue in Ikomia HUB view
- ...
v0.9.1
Improvements:
- Update Python algorithm template
- Update iconography according to new Ikomia branding
- Manage overlay mask while exporting image or video results (#17)
Bug fixes:
v0.9.0
New features:
- OpenCV 4.7.0
- MLFlow 1.30.0
- Remove mandatory authentication for algorithm installation from Ikomia HUB
- Add OpenCV Blur and StackBlur algorithms
- Add keypoints I/O
- Add implicit conversion from instance segmentation I/O to semantic segmentation I/O
Improvements:
- Filtering process on object detection, instance segmentation and semantic segmentation tasks
- System to find best video writer back-end and codec while exporting videos
Bug fixes:
v0.8.1
Improvements:
- Set max fps value for output videos to 25
- Set video data info in workflow video inputs
- Allow to load .webp images
- Improve logging message format
Bug fixes:
v0.8.0
New features:
- New algorithms from the Ikomia HUB: YoloV7, Open MMlab object detection, SparseInst (training and inference)
- Add filtering tasks for object detection, instance segmentation and semantic segmentation
- Add display for text data (Python data dict output for example)
- Add feature to blacklist Python packages that confict with Ikomia built-in packages
- Manage new workflow I/O: object detection, instance segmentation, semantic segmentation
- Add automatic I/O conversion between different types (ex: object detection output -> graphics input)
Improvements:
- Manage compilation architecture for C++ algorithms from Ikomia HUB
- Check Python environment for new user on the same machine
- Export only selected video output on workflow execution
- Allow to load .mxg videos
Bug fixes:
- Disable Tensorboard auto-start to avoid algorithm installation failure
- Crash when adding new workflow input and no data is loaded in the current project
- Display artefact in the plugin manager module when algorithm metadata are too long
- Crashes produced by missing GIL locks
- ...
v0.7.0
New features:
- JupyterLab integration (facilitate plugin code editing)
- New algorithms from the Marketplace (timm framework, mmlab text detection/recognition)
Improvements:
- New video processing system (not restricted to image modification workflow)
- Full batch processing support for videos
- Add support for .ts video format
- Update QuickStart tutorial
- Display both plot and table for numeric outputs
- Better style while burning text to image
- Add timeout support for video reading
Bug fixes:
- Handle invalid full text search keywords for the process library
- Multi-inputs workflow loading
- Multi-threading issues in the video management system
- Unsupported image formats while wrinting videos
- ...
v0.6.1
Improvements:
- Possibility to display multiple tables of results
- Graphics export on videos and streams
- Add support for mxf video format
Bug fixes:
- Incomplete display in table of detection algorithms outputs when several custom measures are computed
- Freeze while exporting video with graphics
- Video export error when output frame sizes are different from source
- Missing GIL protection
- Add trick to fix issue: opencv/opencv#20566
- Open public IP camera from URL
- ...